Closed expression of the interaction kernel in the Bethe-Salpeter equation for quark-antiqu
- 格式:pdf
- 大小:282.88 KB
- 文档页数:27
a rXiv:h ep-th/02226v224O ct25Closed expression of the interaction kernel in the Bethe-Salpeter equation for quark-antiquark bound states Jun-Chen Su .Center for Theoretical Physics,Physics College,Jilin University,Changchun 130023,People’s Republic of China The interaction kernel in the Bethe-Salpeter equation for quark-antiquark bound states is derived from QCD in the general case where the quark and the antiquark can be of different flavors or the same flavor.The kernel is derived with the aid of the Bethe-Salpeter equations satisfied by the quark-antiquark four-point Green function.The latter equations are established based on the equations of motion obeyed by the quark and antiquark propagators,the quark-antiquark four-point Green function and some other kinds of Green functions in which the gluon field is involved.The interaction kernel derived is given a closed and explicit expression.Since the kernel is represented in terms of only a few-types of Green functions,it is not only convenient for perturbative calculations,but also suitable for nonperturbative investigations.PACS:11.10St,12.38.Aw, Key words:Bethe-Salpeter equation,interaction kernel,quark-antiquark bound state.I.INTRODUCTION The Bethe-Salpeter (B-S)equation which was established early in Refs.[1,2]is recognized as a rigorous approach to the relativistic bound state problem and has extensively been investigated in the period of more than half century [3-21].The distinctive features of the equation are:(1)The equation is derived from quantum field theory and hence set up on the firm dynamical basis;(2)The interaction kernel in the equation contains all the interactions taking place in the bound states and therefore the equation provides a possibility of exactly solving the problem of relativistic bound states;(3)The equation is elegantly formulated in a manifestly Lorentz-covariant form in the Minkowski space which allows us to discuss the equation in any coordinate frame.However,there are tremendous difficulties in practical applications of the equation.One of the difficulties arises from the fact that the kernel in the equation was not given a closed form in the ually,the kernel is defined as a sum of all B-S irreducible (or say,two-particle-irreducible)graphs.According to this definition,the kernel can only be calculated by the perturbation method.In practical applications,the perturbation series of the kernel has to be truncated in a ladder approximation.The ladder approximation has been proved to be successful in QED for studying the bound states formed by the electromagnetic interaction [22-26].Nevertheless,it is not feasible in QCD for exploring multi-quark bound states because the confining force which must be taken into account in this case could not follow from a perturbative calculation.It was remarked in Ref.[27]that ”The approach using the Bethe-Salpeter equation has not led to a real breakthrough in our understanding of quark-quark force”.The reason is mainly due to that”we have no method for computing the kernel of the B-S equation”because we have not known the closed expression of the kernel which can be used to evaluate the confining force.In the previous application of the B-S equation to investigate the hadron structure,a phenomenological confining potential is necessarily introduced and added to the one-gluonexchange kernel so as to obtain reasonable theoretical results [27-29].The confining potential was originally proposed from a nonperturbative computation,for example,the lattice gauge calculation for a special Wilson loop [30-33].Since the B-S equation is an exact formalism for bound states and the B-S kernel contains all the interactions responsible for the formation of the bound states,the B-S kernel is supposed to be the most suitable starting point of examining the quark confinement.The question now is concentrated on whether a closed expression of the B-S kernel exists and can be derived?The aim of this paper is to give a definite answer to this question.It will be shown that opposite to the conventional concept that ”The kernel K can not be given in a closed form expression”[34],a closed expression of the B-S kernel for quark-antiquark bound states will perfectly be derived in this paper.The method of deriving the expression of the kernel is usage of various equations of motion satisfied by the qAppendix,the equations of motion satisfied by some Green functions which are necessary to use in the derivation of the B-S kernel are derived.II.DEFINITION OF THE B-S KERNELTo derive the B-S interaction kernel appearing in the B-S equation for qq four-point Green function.For the qψρ(y1)ψ(x)andψT(x),S∗F(y1−y2)ρσ(2.4) whereG(x1,x2;y1,y2)αβρσ= 0+ T{ψα(x1)ψcβ(x2)ψcσ(y2)} 0− (2.5) is the conventional qi 0+ T[ψα(x1)ψcβ(x2)] 0−=S F(x1−x2)αγ(C−1)γβ=S c F(x2−x1)βλCλα(2.6)andi 0+ T{ψcσ(y2)} 0−=CστS F(y2−y1)τρ=(C−1)ρδS c F(y1−y2)δσ(2.7) in whichS F(x1−x2)αγ=1ψγ(x2)] 0− (2.8)andS c F(y1−y2)δσ=1ψcσ(y2)} 0− (2.9)are the ordinary quark and antiquark propagators respectively[5].It is clear that the propagators defined in Eqs.(2.6)and(2.7)are nonzero only for such a quark and an antiquark that they are of the sameflavor.For the quark and antiquark of differentflavors,the Green function defined in Eq.(2.1)is reduced to the conventional form shown in Eq.(2.5)since the second term on the right hand side(RHS)of Eq.(2.4)vanishes.In the case of the quark and antiquark of the sameflavor,the normal product in Eq.(2.1)plays a role of excluding the contraction between the quarkfield and the antiquark one from the Green function.Physically,this avoids the qψ(x)to represent the antiquarkfield in this paper has an advantage that the antiquarkfield would behave as a quark one in the B-S equation so that thequark-antiquark equation formally is the same as the corresponding two-quark equation in the case that the quark and antiquark have differentflavors.Now let us proceed to the derivation of the B-S equation for the qi 0+ T[A aµ(x1)ψγ(x1)Γbν)βλΛc bν(x2|x2,y2)λσ(2.15) hereT b(2.16) withi 0+T[A bν(x2)ψcλ(x2)i0+|T[A aµ(x i)...A bν(x j)A cκ(y k)...A dθ(y l)ψα(x1)i0+|T[A aµ(x i)...A bν(x j)A cκ(y k)...A dθ(y l)ψcβ(x2)i 0+|T[A aµ(x i)...A bν(x j)A cκ(y k)...A dθ(y l)ψα(x1)ψcβ(x2)]|0−(2.20)i 0+|T[A aµ(x i)...A bν(x j)A cκ(y k)...A dθ(y l)ψcσ(y2)]|0−(2.21)G a...bc...dµ...νκ...θ(x i,...,x j;y k,...,y l|x1,x2;y1,y2)αβρσ= 0+|T[A aµ(x i)...A bν(x j)A cκ(y k)...A dθ(y l)ψα(x1)ψcβ(x2)ψcσ(y2)]|0−(2.22)andG a...bc...dµ...νκ...θ(x i,...,x j;y k,...,y l|x1,x2;y1,y2)αβρσ= 0+|T{N[A aµ(x i)...A bν(x j)ψα(x1)ψcβ(x2)]N[ψcσ(y2)A cκ(y k)...A dθ(y l)]}|0−(2.23)where i,j.k,l=1,2.The normal products in Eq.(2.23)are defined in the same way as that in Eq.(2.3).Hereafter, we no longer write individual representations of the Green functions encountered in later derivations.They can be read offfrom the above expressions.According to the derivations given in the appendices,we can write from Eqs.(A24)and(B4)the equations of motion obeyed by the Green function defined in Eq.(2.5)[5][(i∂x1−m1+Σ)G]αβρσ(x1,x2;y1,y2)=δαρδ4(x1−y1)S c F(x2−y2)βσ+Cαβδ4(x1−x2)S∗F(y1−y2)ρσ−(S∗F(y1−y2)ρσ(2.26)where i=1,2.On substituting the definitions in Eqs.(2.4)and(2.26)into Eqs.(2.24)and(2.25),employing the equations in Eqs.(2.10)and(2.14)and the relations in Eqs.(2.6),(2.7),(2.11)and(2.15),it is not difficult to derive the following equations[(i∂x1−m1+Σ)G]αβρσ(x1,x2;y1,y2)=δαρδ4(x1−y1)S c F(x2−y2)βσ−(Γaµ)αγG aµ(x1|x1,x2;y1,y2)γβρσ+ d4z1Σ(x1,z1)αγG(z1,x2;y1,y2)γβρσ(2.27) and[(i∂x2−m2+Σc)G]αβρσ(x1,x2;y1,y2)=δβσδ4(x2−y2)S F(x1−y1)αρ−(Γbν)βλ[(i∂x1−m1+Σ)G bν]αλρσ(x2|x1,x2;y1,y2)+ d4z2Σc(x2,z2)βλ[(i∂x1−m1+Σ)G]αλρσ(x1,z2;y1,y2).(2.29)Thefirst and third terms in the above can directly be evaluated by employing Eqs.(2.10)and(2.27).To compute the second term,wefirst write down the equation of motion for the Green function G bν(x2|x1,x2;y1,y2)which is given by supplementing the self-energy-related term to the both sides of Eq.(A25)[(i∂x1−m1+Σ)G bν]αλρσ(x2|x1,x2;y1,y2)=δαρδ4(x1−y1)Λc bν(x2|x2,y2)λσ+Cαλδ4(x1−x2)Λ∗bν(x2|y1,y2)ρσand G abµν(x1,x2|x1,x2;y1,y2)γλρσare the Green functions which can be read offfrom Eqs.(2.19),(2.21)and(2.22)respectively.According to the relation shown in Eq.(2.26)and the following relation which follows from Eq.(2.23)G abµν(x1,x2|x1,x2;y1,y2)αβρσ=G abµν(x1,x2|x1,x2;y1,y2)αβρσ+Λ∗abµν(x1,x2|x1,x2)αβΛ∗bν(x2|y1,y2)ρσ−(Γaµ)αγG abµν(x1,x2|x1,x2;y1,y2)γλρσ+ d4z1Σ(x1,z1)αγG bν(x2|z1,x2;y1,y2)γλρσ.(2.33) Upon inserting Eqs.(2.10),(2.27)and(2.33)into Eq.(2.29),we arrive at[(i∂x1−m1+Σ)(i∂x2−m2+Σc)G]αβρσ(x1,x2;y1,y2)=δαρδβσδ4(x1−y1)δ4(x2−y2)+H1(x1,x2;y1,y2)αβρσ(2.34) whereH1(x1,x2;y1,y2)αβρσ=5i=1H(i)1(x1,x2;y1,y2)αβρσ(2.35)in whichH(1)1(x1,x2;y1,y2)αβρσ=−(Γaµ)αγ d4z2Σc(x2,z2)βλG aµ(x1|x1,z2;y1,y2)γλρσ,(2.36)H(2)1(x1,x2;y1,y2)αβρσ=−(Γbν)βλG abµν(x1,x2|x1,x2;y1,y2)γλρσ,(2.38)H(4)1(x1,x2;y1,y2)αβρσ=δ4(x1−x2)(Λ∗bν(x2|y1,y2)ρσ(2.39) andH(5)1(x1,x2;y1,y2)αβρσ= d4z1d4z2Σ(x1,z1)αγΣc(x2,z2)βλG(z1,z2;y1,y2)γλρσ.(2.40)From perturbative calculations or irreducible decompositions of the Green functions[1-5,21],it may be found that the function H1(x1,x2;y1,y2)is B-S reducible and can be written in the formH 1(x 1,x 2;y 1,y 2)αβρσ= d 4z 1d 4z 2K (x 1,x 2;z 1,z 2)αβγλG (z 1,z 2;y 1,y 2)γλρσ(2.41)where K (x 1,x 2;z 1,z 2)is precisely the B-S irreducible kernel.With the above expression for the function H 1(x 1,x 2;y 1,y 2),the equation in Eq.(2.34)can be expressed in a closed form[(i∂x 1−m 1+Σ)(i∂x 2−m 2+Σc )G ]αβρσ(x 1,x 2;y 1,y 2)=δαρδβσδ4(x 1−y 1)δ4(x 2−y 2)+ d 4z 1d 4z 2K (x 1,x 2;z 1,z 2)αβγλG (z 1,z 2;y 1,y 2)γλρσ.(2.42)This just is the B-S equation satisfied by the Green function G (x 1,x 2;y 1,y 2).By making use of the Lehmann rep-resentation of the Green function G (x 1,x 2;y 1,y 2)or the well-known procedure proposed by Gell-Mann and Low [2],one may readily derive from Eq.(2.42)the B-S equation satisfied by B-S amplitudes describing the qq bound state and ςmarks the otherquantum numbers of the state.The above equation can be written in the form of an integral equation when we operate on the both sides of the above equation with (i∂x 1−m 1+Σ)−1(i∂x 2−m 2+Σc )−1[1-5]χP ς(x 1,x 2)= d 4z 1d 4z 2d 4y 1d 4y 2S F (x 1−z 1)S c F (x 2−z 2)K (z 1,z 2;y 1,y 2)χP ς(y 1,y 2)(2.45)III.DERIV ATION OF THE B-S KERNELBeyond the perturbation method,the B-S kernel may be derived by starting from its definition shown in Eq.(2.41).One method of the derivation is usage of the technique of irreducible decomposition of Green functions.By this method,once the one and two-particle-irreducible decompositions of the Green functions contained in the function H 1(x 1,x 2;y 1,y 2)are completed,it can be found that the function H 1(x 1,x 2;y 1,y 2)can really be represented in the form as written in Eq.(2.41)and,at the same time,the kernel will be explicitly worked out.Nevertheless,the expression of the kernel given in this way is much complicated (We leave the detailed discussions in the future).Another method of deriving the kernel is to use B-S equations which describe variations of the Green functions involved in the function H 1(x 1,x 2;y 1,y 2)with the coordinates y 1and y 2.In this paper,we adopt the latter method and describe the procedure of the derivation in this section.Let us operate on the both sides of Eq.(2.41)from the right with the operator (i ←−∂y 1+m 1−Σ)(i ←−∂y 2+m 2−Σc )such thatd 4z 1d 4z 2K (x 1,x 2;z 1,z 2)αβγλ[G (i ←−∂y 1+m 1−Σ)(i ←−∂y 2+m 2−Σc )]γλρσ(z 1,z 2;y 1,y 2)=Q (x 1,x 2;y 1,y 2)αβρσ(3.1)whereQ (x 1,x 2;y 1,y 2)αβρσ=[H 1(i ←−∂y 1+m 1−Σ)(i ←−∂y 2+m 2−Σc )]αβρσ(x 1,x 2;y 1,y 2).(3.2)As seen from Eqs.(3.1)and (3.2),to derive the B-S kernel,it is necessary to use the B-S equations with re-spect to y 1and y 2for the Green function G (x 1,x 2;y 1,y 2)and the other Green functions appearing in the function H 1(x 1,x 2;y 1,y 2).Analogous to the procedure stated in the preceding section,the above-mentioned B-S equations can be derived with the aid of equations of motion with respect to y 1and y 2for those Green functions.First,we show the equations of motion obeyed by the quark and antiquark propagators[S F (i ←−∂y 1+m 1−Σ)]αρ(x 1,y 1)=−δαρδ4(x 1−y 1)(3.3)and[S c F (i ←−∂y 2+m 2−Σc )]βσ(x 2,y 2)=−δβσδ4(x 2−y 2)(3.4)which are given by Eqs.(C2)and (D2)respectively with the source J being set to vanish.We note here that the self-energies,as easily proved,are the same as those defined in Eqs.(2.11)and (2.15).Next,we derive the B-S equation satisfied by the Green function G (x 1,x 2;y 1,y 2).For this purpose,we first write down the equation of motion describing variation of the conventional Green function G (x 1,x 2;y 1,y 2)with y 1which is obtained from Eq.(C5)by setting J =0and adding the term related to the self-energy[G (i ←−∂y 1+m 1−Σ)]αβρσ(x 1,x 2;y 1,y 2)=−δαρδ4(x 1−y 1)S c F (x 2−y 2)βσ−C ρσδ4(y 1−y 2)S ∗F (x 1−x 2)αβ+G a µ(y 1|x 1,x 2;y 1,y 2)αβτσ(Γaµ)τρ− d 4z 1G (x 1,x 2;z 1,y 2)αβτσΣ(z 1,y 1)τρ(3.5)where G a µ(y 1|x 1,x 2;y 1,y 2)is the Green function with a gluon field at y 1in it whose expression can be read from Eq.(2.22).Substituting in Eq.(3.5)the relation given in Eq.(2.4)and the following relation which follows from Eq.(2.23)G a µ(y i |x 1,x 2;y 1,y 2)αβρσ=G a µ(y i |x 1,x 2;y 1,y 2)αβρσ+S ∗F (x 1−x 2)αβS ∗F (i ←−∂y 1+m 1−Σ)]ρσ(y 1,y 2)=C ρσδ4(y 1−y 2)(3.7)which is given by Eqs.(C8)-(C10),it can be found[G (i ←−∂y 1+m 1−Σ)]αβρσ(x 1,x 2;y 1,y 2)=−δαρδ4(x 1−y 1)S c F (x 2−y 2)βσ+G a µ(y 1|x 1,x 2;y 1,y 2)αβτσ(Γaµ)τρ− d 4z 1G (x 1,x 2;z 1,y 2)αβτσΣ(z 1,y 1)σρ.(3.8)In order to derive the B-S equation,we need to operate on the both sides of the above equation with the operator (i ←−∂y 2+m 2−Σc ),[G (i ←−∂y 1+m 1−Σ)(i ←−∂y 2+m 2−Σc )]αβρσ(x 1,x 2;y 1,y 2)=δαρδβσδ4(x 1−y 1)δ4(x 2−y 2)+[G a µ(i ←−∂y 2+m 2−Σc )]αβτσ(y 1|x 1,x 2;y 1,y 2)(Γaµ)τρ− d 4z 1[G (i ←−∂y 2+m 2−Σc )]αβτσ(x 1,x 2;z 1,y 2)Σ(z 1,y 1)τρ(3.9)where Eq.(3.4)has been used.For computing the third term in the equation,we start from the equation of motion which is obtained from Eq.(D8)with setting J =0and supplementing the self-energy-related term[G (i ←−∂y 2+m 2−Σc )]αβρσ(x 1,x 2;y 1,y 2)=−δβσδ4(x 2−y 2)S F (x 1−y 1)αρ−C ρσδ4(y 1−y 2)S ∗F (x 1−x 2)αβ+G b ν(y 2|x 1,x 2;y 1,y 2)αβρδ(S ∗F (i ←−∂y 2+m 2−Σc )]ρσ(y 1,y 2)=C ρσδ4(y 1−y 2)(3.11)which is given by Eqs.(D14)-(D16)and the relations represented in Eq.(2.15),one can get such an equation that[G (i ←−∂y 2+m 2−Σc )]αβρσ(x 1,x 2;y 1,y 2)=−δβσδ4(x 2−y 2)S F (x 1−y 1)αρ+G b ν(y 2|x 1,x 2;y 1,y 2)αβρδ(Λ∗a µ(i ←−∂y 2+m 2−Σc )]ρσ(y 1|y 1,y 2).(3.13)The first term can be calculated by virtue of the equation of motion derived in Eq.(D9).With addition of the self-energy-relevant term to Eq.(D9),we have[G aµ(i ←−∂y2+m2−Σc)]αβρσ(y1|x1,x2;y1,y2)=−δβσδ4(x2−y2)Λaµ(y1|x1,y1)αρ−Cρσδ4(y1−y2)Λ∗aµ(y1|x1,x2)αβ+G abµν(y1,y2|x1,x2;y1,y2)αβρδ(Λ∗ab µν(y1,y2|y1,y2)ρσ(3.15)whereΓbν)δσ− d4z2G aµ(y1|x1,x2;y1,z2)αβρδΣc(z2,y2)δσ−S∗F(x1−x2)αβΓbν)δσ+S∗F(x1−x2)αβ d4z2Λ∗aµ(i ←−∂y2+m2−Σc)]ρσ(y1|y1,y2)=Γbν)δσ− d4z2Γbν)δσ− d4z2G aµ(y1|x1,x2;y1,z2)αβρδΣc(z2,y2)δσ.(3.18) Substitution of Eqs.(3.12)and(3.18)in Eq.(3.9)yields[G(i ←−∂y1+m1−Σ)(i←−∂y2+m2−Σc)]αβρσ(x1,x2;y1,y2)=δαρδβσδ4(x1−y1)δ4(x2−y2)+H2(x1,x2;y1,y2)αβρσ(3.19) whereH2(x1,x2;y1,y2)αβρσ=5j=1H(j)2(x1,x2;y1,y2)αβρσ(3.20)in whichH(1)2(x1,x2;y1,y2)αβρσ=− d4z2G aµ(y1|x1,x2;y1,z2)αβτδΣc(z2,y2)δσ(Γaµ)τρ,(3.21)H(2)2(x1,x2;y1,y2)αβρσ=− d4z1G bν(y2|x1,x2;z1,y2)αβτδΣ(z1,y1)τρ(Γbν)δσ,(3.23)H(4)2(x1,x2;y1,y2)αβρσ=δ4(y1−y2)Λ∗aµ(y1|x1,x2)αβ(CΓaµ)σρ(3.24) andH(5)2(x1,x2;y1,y2)αβρσ= d4z1d4z2G(x1,x2;z1,z2)αβτδΣ(z1,y1)τρΣc(z2,y2)δσ.(3.25) By applying the B-S equation given in Eq.(3.19),Eq.(3.1)can be written asK(x1,x2;y1,y2)αβρσ=Q(x1,x2;y1,y2)αβρσ− d4z1d4z2K(x1,x2;z1,z2)αβτδH2(z1,z2;y1,y2)τδρσ.(3.26) To reach a closed expression of the B-S kernel,it is necessary to eliminate the kernel appearing in the second term on the RHS of Eq.(3.26).For this purpose,we operate on the both sides of Eq.(2.41)with the inverse of the Green function G(x1,x2;y1,y2).With this operation,noticingd4z1d4z2G(x1,x2;z1,z2)αβγλG−1(z1,z2;y1,y2)γλρσ=δαρδβσδ4(x1−y1)δ4(x2−y2),(3.27) one getsK(x1,x2;z1,z2)αβτδ= d4u1d4u2H1(x1,x2;u1,u2)αβγλG−1(u1,u2;z1,z2)γλτδ.(3.28) Upon inserting the above expression into Eq.(3.26),wefinally arrive atK(x1,x2;y1,y2)αβρσ=Q(x1,x2;y1,y2)αβρσ−S(x1,x2;y1,y2)αβρσ(3.29) whereS(x1,x2;y1,y2)αβρσ= d4u1d4u2d4v1d4v2H1(x1,x2;u1,u2)αβγλG−1(u1,u2;v1,v2)γλτδH2(v1,v2;y1,y2)τδρσ.(3.30) As we see,the B-S kernel shown above contains two terms.The second term has been explicitly written out.It contains a few types of the Green functions as well as the self-energies appearing in the mutually conjugate functions H1(x1,x2;y1,y2)and H2(x1,x2;y1,y2)shown in Eqs.(2.35)-(2.40)and(3.20)-(3.25)respectively.Clearly,to give afinal expression of the B-S kernel,according to the definition in Eq.(3.2),we need to calculate the function Q(x1,x2;y1,y2)by means of the B-S equations with respect to y1and y2for the Green functions involved in the function H1(x1,x2;y1,y2).We leave the calculations in the next section.IV.CLOSED EXPRESSION OF THE B-S KERNELThis section is used to calculate the function Q(x1,x2;y1,y2)so as to give thefinal expression of the B-S kernel. In accordance with the expressions shown in Eqs.(3.2)and(2.35),we can writeQ(x1,x2;y1,y2)αβρσ=5i=1Q i(x1,x2;y1,y2)αβρσ(4.1)whereQ i(x1,x2;y1,y2)αβρσ=[H(i)1(i ←−∂y1+m1−Σ)(i←−∂y2+m2−Σc)]αβρσ(x1,x2;y1,y2).(4.2)This expression indicates that to compute each Q i(x1,x2;y1,y2),we need to derive the B-S equations with respect to y1and y2for the Green functions contained in the function H1(x1,x2;y1,y2).We will separately calculate each Q i(x1,x2;y1,y2)below.a.Calculation of Q1(x1,x2;y1,y2)Based on Eqs.(4.2)and(2.36),we can writeQ1(x1,x2;y1,y2)αβρσ=−(Γaµ)αγ d4u2Σc(x2,u2)βλ[G aµ(i←−∂y1+m1−Σ)(i←−∂y2+m2−Σc)]γλρσ(x1|x1,u2;y1,y2).(4.3) To calculate the above function,it is necessary to start from the following equation of motion which is established from Eq.(C6)by setting the source J=0,G aµ(x1|x1,u2;y1,y2)γλτσ(i ←−∂y1+m1)τρ=−δγρδ4(x1−y1)Λc aµ(x1|u2,y2)λσ−Cρσδ4(y1−y2)Λ∗aµ(x1|x1,u2)γλ+G abµν(x1;y1|x1,u2;y1,y2)γλτσ(Γaµ)τρ(4.4)where the Green functions have been represented respectively in Eqs.(2.19),(2.20)and(2.22).Considering the definitions given in Eq.(2.26)and in the followingG abµν(x i;y j|x1,u2;y1,y2)γλρσ=G abµν(x i;y j|x1,u2;y1,y2)γλρσ+Λ∗aµ(x i|x1,u2)γλΓbν)δσ− d4v2Λcaµ(x1|u2,v2)λδΣc(v2,y2)δσ(4.8) where the two Green functions were represented in Eq.(2.19).The third term in Eq.(4.7)can be calculated by virtue of the equation given in Eq.(D10)G aµ(x1|x1,u2;v1,y2)γλτδ(i ←−∂y2+m2)δσ=−δλσδ4(u2−y2)Λaµ(x1|x1,v1)γτ−Cτσδ4(v1−y2)Λ∗aµ(x1|x1,u2)γλ+G abµν(x1;y2|x1,u2;v1,y2)γλτδ(Γbν)δσ− d4v2G aµ(x1|x1,u2;v1,v2)γλτδΣc(v2,y2)δσ.(4.10) For calculating the second term in Eq.(4.7),it is needed to use the following equation obtained from Eq.(D11)by setting J=0[G abµν(i ←−∂y2+m2)]γλτσ(x1;y1|x1,u2;y1,y2)γλτσ=−δλσδ4(u2−y2)Λabµν(x1;y1|x1,y1)γτ−Cτσδ4(y1−y2)Λ∗abµν(x1;y1|x1,u2)γλ+G abcµνκ(x1;y1,y2|x1,u2;y1,y2)γλτδ(Λ∗bcνκ(y1,y2|y1,y2)τσ(4.12)which is implied by Eq.(2.23)and utilizing Eqs.(4.11)and(3.17),an equation of motion for the Green function G abµν(x1;y1|x1,u2;y1,y2)will be found to be[G abµν(i ←−∂y2+m2−Σc)]γλτσ(x1;y1|x1,u2;y1,y2)γλτσ=−δλσδ4(u2−y2)Λabµν(x1;y1|x1,y1)γτ−Cτσδ4(y1−y2)Λ∗abµν(x1;y1|x1,u2)γλ+G abcµνκ(x1;y1,y2|x1,u2;y1,y2)γλτδ(When Eqs.(4.8),(4.10)and(4.13)are inserted into Eq.(4.7)and then Eq.(4.7)is inserted into Eq.(4.3),an explicit expression of the function Q1(x1,x2;y1,y2)will be derived as given in the followingQ1(x1,x2;y1,y2)αβρσ=R1(x1,x2;y1,y2)αβρσ+M1(x1,x2;y1,y2)αβρσ+ d4v1d4v2H(1)1(x1,x2;v1,v2)αβτδΣ(v1,y1)τρΣc(v2,y2)δσ(4.14) whereR1(x1,x2;y1,y2)αβρσ=δ4(x1−y1)(Γaµ)αρ d4u2Σc(x2,u2)βλ[Λc abµν(x1;y2|u2,y2)λδ(Γcκ)δσ− d4v2G abµν(x1;y1|x1,u2;y1,v2)γλτδΣc(v2,y2)δσ(Γbν)τρ− d4v1G abµν(x1;y2|x1,u2;v1,y2)γλτδΣ(v1,y1)τρ(Γaµ)βλ d4u1Σ(x1,u1)αγ[G aµ(i←−∂y1+m1−Σ)(i←−∂y2+m2−Σc)]γλρσ(x2|u1,x2;y1,y2)(4.17) In comparison of the above expression with that denoted in Eq.(4.3),it is clearly seen that the function Q2(x1,x2;y1,y2)αβρσcan explicitly be written out from the expression of Q1(x1,x2;y1,y2)αβρσshown in Eqs.(4.14)-(4.16)by the following replacements:(Γaµ)αγ d4u2Σc(x2,u2)βλ→(Γaµ)βσ d4u1Σ(x1,u1)αγ[Λabµν(x2;y1|u1,y1)γτ(Γbν)τρ− d4v1Λaµ(x2|u1,v1)γτΣ(v1,y1)τρ]+Σ(x1,y1)αρ(Γbν)δσ− d4v2Λc aµ(x2|x2,v2)λδΣc(v2,y2)δσ]−δ4(y1−y2)(Γaµ)βλ d4u1Σ(x1,u1)αγ[G abcµνκ(x2;y1,y2|u1,x2;y1,y2)γλτδ(Γbν)τρ((4.20)Γbν)δσ]and the function H(2)1(x1,x2;v1,v2)was defined in Eq.(2.37).c.Calculation of Q3(x1,x2;y1,y2)In accordance with Eqs.(4.2)and(2.38),the function Q3(x1,x2;y1,y2)is known to beQ3(x1,x2;y1,y2)αβρσ=(Γaµ)αγ(G abµν(x1,x2|x1,x2;y1,y2)γλτσ(i ←−∂y1+m1)τρ=−δγρδ4(x1−y1)Λc abµν(x1,x2|x2,y2)λσ−Cρσδ4(y1−y2)Λ∗abµν(x1,x2|x1,x2)γλ+G abcµνκ(x1,x2;y1|x1,x2;y1,y2)γλτσ(Γcκ)τρ(4.22)where the Green functions were respectively represented in Eqs.(2.19),(2.20)and(2.22).Grounded on the definition given in Eqs.(2.31)and the following one which can be written out from Eq.(2.23):G abc µνκ(x1,x2,y i|x1,x2;y1,y2)γλτσ=G abcµνκ(x1,x2,y i|x1,x2;y1,y2)γλτσ+Λ∗abµν(x1,x2|x1,x2)γλΓcκ)δσ− d4v2Λc abµν(x1,x2|x2,v2)λδΣc(v2−y2)δσ(4.26) wherei∆abµν(x1−x2)= 0+ T[A aµ(x1)A bν(x2)] 0− (4.27) is the gluon propagator and the other Green functions were defined in Eq.(2.19).The third term in Eq.(4.25)will be evaluated with the aid of the equation given by Eq.(D12)G abµν(x1,x2|x1,x2;v1,y2)γλτδ(i ←−∂y2+m2)δσ=−δλσδ4(x2−y2)Λabµν(x1,x2|x1,v1)γτ−Cτσδ4(v1−y2)Λ∗abµν(x1,x2|x1,x2)γλ+G abcµνκ(x1,x2;y2|x1,x2;v1,y2)γλτδ(Γcκ)δσ− d4v2G abµν(x1,x2|x1,x2;v1,v2)γλτδΣc(v2,y2)δσ.(4.29)The second term in Eq.(4.25)can be calculated by means of the equation derived in Eq.(D13)G abcµνκ(x1,x2;y1|x1,x2;y1,y2)γλτδ(i←−∂y2+m2)δσ=−δλσδ4(x2−y2)Λabcµνκ(x1,x2;y1|x1,y1)γτ−Cτσδ4(y1−y2)Λ∗abcµνκ(x1,x2;y1|x1,x2)γλ+G abcdµνκθ(x1,x2;y1,y2|x1,x2;y1,y2)γλτδ(Λ∗cdκθ(y1,y2|y1,y2)τσ(4.31)which follows from Eq.(2.23)and applying the equations in Eqs.(4.30)and(3.17),it is easy tofind[G abcµνκ(i←−∂y2+m2−Σc)]γλτσ(x1,x2;y1|x1,x2;y1,y2)=−δλσδ4(x2−y2)Λabcµνκ(x1,x2;y1|x1,y1)γτ−Cτσδ4(y1−y2)Λ∗abcµνκ(x1,x2;y1|x1,x2)γλ+G abcdµνκθ(x1,x2;y1,y2|x1,x2;y1,y2)γλτδ(Upon substituting Eqs.(4.26),(4.29)and(4.32)in Eq.(4.25)and then inserting Eq.(4.25)into Eq.(4.21),we eventually arrive atQ3(x1,x2;y1,y2)αβρσ=K(0)t(x1,x2;y1,y2)αβρσ+Γaν)βσi∆abµν(x1−x2)(4.34) which is the t-channel one-gluon exchange kernel,Γaν)βλG abcdµνκθ(x1,x2;y1,y2|x1,x2;y1,y2)γλτδ(Γcκ)τρ(Γaν)βλ[Λc abcµνκ(x1,x2;y2|x2,y2)λδ(Γbν)βσ((Γaµ)αγ[Λabcµνκ(x1,x2;y1|x1,y1)γτ(Γcκ)τρ− d4v1Λabµν(x1,x2|x1,v1)γτΣ(v1,y1)τρ]+δ4(y1−y2)(Γaµ)αγ(Γaν)βλ[ d4v2G abcµνκ(x1,x2;y1|x1,x2;y1,v2)γλτδΣc(v2,y2)δσ(Γcκ)τρ+ d4v1G abcµνκ(x1,x2;y2|x1,x2;v1,y2)γλτδΣ(v1,y1)τρ(ΓaµC)βα[Λ∗aµ(i ←−∂y1+m1−Σ)]ρσ(x2|y1,y2)=Λ∗aµ(x2|v1,y2)τσΣ(v1,y1)τρ(4.39)where the Green functions was written in Eq.(2.21). Now,let us operate on Eq.(4.39)with the operator(i ←−∂y2+m2−Σc),[Λ∗ab µν(i←−∂y2+m2−Σc)]τσ(x2;y1|y1,y2)(Γbν)τρ− d4v1[Λ∗aµ(i ←−∂y2+m2−Σc)]τσ(x2|v1,y2)=Γbν)δσ− d4v2Λ∗ab µν(i←−∂y2+m2−Σc)]τσ(x2,y1|y1,y2)=Cτσδ4(y1−y2)i∆abµν(x2−y1)+Γcκ)δσ− d4v2Substituting atfirst Eqs.(4.41)and(4.42)into Eq.(4.40)and then Eq.(4.40)into Eq.(4.38),the function Q4(x1,x2;y1,y2)will be explicitly written in the formQ4(x1,x2;y1,y2)αβρσ=K(0)s(x1,x2;y1,y2)αβρσ+R4(x1,x2;y1,y2)αβρσ+ d4v1d4v2H(4)1(x1,x2;v1,v2)αβτδΣ(v1,y1)τρΣc(v2,y2)(4.43) whereK(0)s(x1,x2;y1,y2)αβρσ=δ4(x1−x2)δ4(y1−y2)(ΓaµC)βα{[Γcκ)δσ− d4v2Λ∗abµν(x2;y2|v1,y2)τδΣ(v1,y1)τρ(q annihilation and creation process which takes place between the quark(antiquark)in the initial state and the antiquark(quark)in thefinal state as indicated by the gluon propagator in Eq.(4.44),not between the quark and the antiquark both of which belong to the initial state or thefinal state B-S amplitude for a bound state.e.Calculation of Q5(x1,x2;y1,y2)From Eqs.(4.2)and(2.40),we haveQ5(x1,x2;y1,y2)αβρσ= d4u1d4u2Σ(x1,u1)αγΣc(x2,u2)βλ[G(i←−∂y1+m1−Σ)(i←−∂y2+m2−Σc)]γλρσ(u1,u2;y1,y2).(4.46) By applying the equation given in Eq.(3.19),the above function can explicitly be represented asQ5(x1,x2;y1,y2)αβρσ=R0(x1,x2;y1,y2)αβρσ+ d4u1d4u2Σ(x1,u1)αγΣc(x2,u2)βλH2(u1,u2;y1,y2)γλρσ(4.47) whereR0(x1,x2;y1,y2)αβρσ=Σ(x1,y1)αρΣc(x2,y2)βσ(4.48) and H2(u1,u2;y1,y2)γλρσwas shown in Eqs.(3.20)-(3.25).In summary,combining the expressions given in Eqs.(4.14),(4.18),(4.33),(4.43)and(4.47),an explicit expression of the function Q(x1,x2;y1,y2)written in Eq.(4.1)will be eventually found as shown in the followingQ(x1,x2;y1,y2)αβρσ=K(0)t(x1,x2;y1,y2)αβρσ+K(0)s(x1,x2;y1,y2)αβρσ+here j=1,2and all the functions on the RHS of Eqs.(4.50)-(4.52)have already been explicitly given before.Now let us turn to the function S(x1,x2;y1,y2)αβρσexpressed in Eq.(3.30).According to the definitions denoted in Eqs.(2.35),(3.20)and(4.52),we can writeH j(x1,x2;y1,y2)αβρσ=H j(x1,x2;y1,y2)αβρσ+H(5)j(x1,x2;y1,y2)αβρσ(4.53) .Thus,Eq.(3.30)can be rewritten in the formS(x1,x2;y1,y2)αβρσ= d4u1d4u2d4v1d4v2H1(x1,x2;u1,u2)αβγλG−1(u1,u2;v1,v2)γλτδH2(v1,v2;y1,y2)τδρσ+ d4u1d4u2d4v1d4v2H(5)1(x1,x2;u1,u2)αβγλG−1(u1,u2;v1,v2)γλτδH2(v1,v2;y1,y2)τδρσ+ d4u1d4u2d4v1d4v2H1(x1,x2;u1,u2)αβγλG−1(u1,u2;v1,v2)γλτδH(5)2(v1,v2;y1,y2)τδρσ.(4.54)When the functions H(5)1(x1,x2;u1,u2)αβγλand H(5)2(v1,v2;y1,y2)τδρσin the second and third terms are respectively replaced by the expressions given in Eqs.(2.40)and(3.25)and utilizing the relation in Eq.(3.27),it is easy to see that the last two terms in Eq.(4.54)exactly equal to the last two terms in Eq.(4.49)respectively.Therefore,when we substitute Eqs.(4.49)and(4.54)into Eq.(3.29),the above-mentioned terms are cancelled with each other.Thus, the B-S kernel will befinally expressed byK(x1,x2;y1,y2)αβρσ=Q(x1,x2;y1,y2)αβρσ−S(x1,x2;y1,y2)αβρσ(4.55) whereQ(x1,x2;y1,y2)αβρσ=K(0)t(x1,x2;y1,y2)αβρσ+K(0)s(x1,x2;y1,y2)αβρσ+q Green functions and given a closed and explicit expression shown in Eqs.(4.55)-(4.57).This expression is manifestly Lorentz-covariant even in the space-like region of Minkowski space where the bound states exist and it suits to not only the case that the quark and antiquark have differentflavors,but also the case where they have the sameflavor. In the latter case,the masses of the quark and antiquark are the same,m1=m2=m.Meanwhile,the B-S kernel contains the part describing the qΛ∗a...b...µ...ν...(x i...,y j,...|y1,y2)which give the higher order perturbative corrections to the vertices of the s-channel one-gluon exchange kernel.It is well known that the s-channel one-gluon exchange kernel gives no contribution to the q q color singlet.But, this does not imply that the qq annihilation part of the kernel is absent.In this case,the B-S kernel is of a simpler expression which can be written out from the corresponding expression given in the preceding section by deleting out the q。